Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Codebase-Memory: Tree-Sitter-Based Knowledge Graphs for LLM Code Exploration via MCP

Martin Vogel, Falk Meyer-Eschenbach, Severin Kohler, Elias Grünewald, Felix Balzer

Abstract

Large Language Model (LLM) coding agents typically explore codebases through repeated file-reading and grep-searching, consuming thousands of tokens per query without structural understanding. We present Codebase-Memory, an open-source system that constructs a persistent, Tree-Sitter-based knowledge graph via the Model Context Protocol (MCP), parsing 66 languages through a multi-phase pipeline with parallel worker pools, call-graph traversal, impact analysis, and community discovery. Evaluated across 31 real-world repositories, Codebase-Memory achieves 83% answer quality versus 92% for a file-exploration agent, at ten times fewer tokens and 2.1 times fewer tool calls. For graph-native queries such as hub detection and caller ranking, it matches or exceeds the explorer on 19 of 31 languages.

Codebase-Memory: Tree-Sitter-Based Knowledge Graphs for LLM Code Exploration via MCP

Abstract

Large Language Model (LLM) coding agents typically explore codebases through repeated file-reading and grep-searching, consuming thousands of tokens per query without structural understanding. We present Codebase-Memory, an open-source system that constructs a persistent, Tree-Sitter-based knowledge graph via the Model Context Protocol (MCP), parsing 66 languages through a multi-phase pipeline with parallel worker pools, call-graph traversal, impact analysis, and community discovery. Evaluated across 31 real-world repositories, Codebase-Memory achieves 83% answer quality versus 92% for a file-exploration agent, at ten times fewer tokens and 2.1 times fewer tool calls. For graph-native queries such as hub detection and caller ranking, it matches or exceeds the explorer on 19 of 31 languages.

Paper Structure

This paper contains 31 sections, 1 figure, 8 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Structural Code Analysis
  4. Code Retrieval for LLMs
  5. LLM Coding Agents
  6. Token Efficiency
  7. System Design
  8. Architecture Overview
  9. Graph Schema
  10. Multi-Pass Pipeline
  11. Call Resolution
  12. LSP-Style Hybrid Type Resolution.
  13. MCP Tool Interface
  14. Incremental Synchronization
  15. Community Detection
  16. ...and 16 more sections

Figures (1)

  • Figure 1: Codebase-Memory architecture. Source files are parsed via Tree-Sitter, stored as a knowledge graph in SQLite, and exposed to LLM agents through 14 typed MCP tools. A file watcher triggers incremental re-indexing on changes.